Learning Transportation Modes from Raw GPS Data

1. A method comprising: obtaining positioning data from a user device of a user, the positioning data identifying positions of the user device over a period of time as determined by a location sensor of the user device; determining a current transportation mode of the user based at least upon the positioning data, the current transportation mode being determined based at least on a transition probability from a previously-determined transportation mode of the user to the current transportation mode; and providing information to the user based at least upon the current transportation mode.

2. The method of claim 1 , the positioning data comprising Global Positioning System (GPS) data and the location sensor comprising a GPS sensor.

3. The method of claim 1 , further comprising: extracting features from the positioning data; and evaluating the features with an inference model to determine the current transportation mode.

4. The method of claim 3 , further comprising: learning the inference model from a dataset of other positioning data of other users.

5. The method of claim 3 , wherein: the positioning data includes location coordinates of the user device and timestamps indicating when the user device was present at the location coordinates, and the extracting the features include extracting, based at least on the location coordinates and the timestamps, a first feature reflecting velocity of the user device and a second feature reflecting acceleration of the user device.

6. The method of claim 1 , the providing comprising: providing first information to the user when the current transportation mode is determined to comprise a first transportation mode; and providing second information to the user when the current transportation mode is determined to comprise a second transportation mode.

7. The method of claim 6 , wherein the first transportation mode is walking, the second transportation mode is riding in a vehicle, the first information includes a map presented at a first scale for the user when walking, and the second information includes the map presented at a second scale for the user when riding in the vehicle, the first scale being different than the second scale.

8. A system comprising: a processing unit; and a volatile or non-volatile storage device storing instructions which, when executed by the processing unit, cause the processing unit to: determine a first transportation mode associated with a first segment based at least upon a second transportation mode associated with a second segment.

9. The system of claim 8 , wherein the first segment and the second segment are determined using positioning data that identifies locations of a device.

10. The system of claim 9 , the first segment corresponding to a first time interval and the second segment corresponding to a second time interval that occurs before the first time interval.

11. The system of claim 10 , wherein the instructions, when executed by the processing unit, cause the processing unit to: determine a probability associated with a transition between the second transportation mode and the first transportation mode, wherein the first transportation mode is determined based at least on the probability.

12. The system of claim 10 , wherein the instructions, when executed by the processing unit, cause the processing unit to: separate the positioning data into the first segment and the second segment based at least on velocity or acceleration data obtained from the positioning data; extract first features from the positioning data by processing first locations and first timestamps that are included in the positioning data and that are associated with the first segment; extract second features from the positioning data by processing second locations and second timestamps that are included in the positioning data and that are associated with the second segment; and determine the first transportation mode based at least on the first features and determine the second transportation mode based at least on the second features.

13. The system of claim 12 , wherein the instructions, when executed by the processing unit, cause the processing unit to: determine first probabilities of different transportation modes for the first segment based at least on the first features; determine second probabilities of transitioning from the second transportation mode to the different transportation modes; modify the first probabilities using corresponding second probabilities; and select the first transportation mode using the modified first probabilities.

14. The system of claim 8 , wherein the first transportation mode or the second transportation mode comprises at least one of walking, driving, bicycling, or commuting by bus.

15. A system comprising: a processing unit; and a volatile or non-volatile storage device storing instructions which, when executed by the processing unit, cause the processing unit to: merge a first segment of positioning data with a second segment of positioning data based at least upon a determination that the first segment has a length below a threshold length.

16. The system of claim 15 , wherein the instructions, when executed by the processing unit, cause the processing unit to: determine a second transportation mode for the second segment based at least upon the positioning data; and apply the second transportation mode to the first segment based at least on the determination that the first segment is below the threshold length.

17. The system of claim 16 , wherein the positioning data identifies positions of a mobile device that generates the positioning data and the threshold length is expressed as a distance.

18. The system of claim 17 , wherein the positioning data indicates that the mobile device traveled the second segment prior the first segment.

19. The system of claim 15 , embodied as a mobile device.

20. The method of claim 1 , the location data comprising Global Positioning System (GPS) data, the method further comprising: processing the GPS data to identify a plurality of walk segments where the user is determined to have been walking and a plurality of non-walk segments where the user is determined to have been using a transportation mode other than walking; and after identifying the plurality of walk segments and the plurality of non-walk segments, classifying the non-walk segments into at least two different vehicular transportation modes.

21. The method of claim 20 , the at least two different vehicular transportation modes comprising travelling in a car and travelling on a bike.

22. The method of claim 20 , the at least two different vehicular transportation modes comprising riding a bus and travelling on a bike.